A carregar...
Publicação
Trees of green : constructing panels of tree canopy from aerial imagery
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 17.36 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This paper develops a fully-automated workflow for
constructing panels of tree canopy from high-resolution multispectral imagery with limited near-infrared (NIR) training data. The
proposed workflow utilizes the tree-pixel detection algorithm developed by Yang, Wu, Praun, and Ma (2009) and Bosch (2020) on a
large set of U.S. urban areas but modifies it by creating automatic
ground-truth masks through various visual graphics techniques
that leverage modern high-resolution NIR data. By matching colors
across different imagery periods, the workflow predicts tree presence
in older images without NIR data, using the recent images with
NIR data. Using a subset of cities that represent the different U.S.
climate regions, I quantify the effectiveness of the workflow by
implementing the algorithm without pre-processing in the creation
of ground-truth masks, without equalizing colors across periods,
and using a universal model for all areas. The comparison shows
that my workflow is the option that leads to better results in terms
of accuracy, recall, and precision.
Descrição
Palavras-chave
aerial imagery tree detection near-infrared light panel data
Contexto Educativo
Citação
Miñano-Mañero, Alba (2024). "Trees of green : constructing panels of tree canopy from aerial imagery". REM Working paper series, nº 0352/2024
Editora
ISEG – REM (Research in Economics and Mathematics)